Hacker News Reader

Comment by plainOldText | Hacker News Reader

> It’s not like every local organization can marshal expertise on every issue.

In order to be efficient decision makers, local organizations should be concerned with acquiring expertise related to local issues, and then, through cooperation, global expertise becomes an emergent property. It's not like every central institution can marshal expertise on every local issue.

One reason you want to avoid centralization is due to the fact that decision making occurs under imperfect and incomplete knowledge, or bounded-rationality as Herbert Simon called it.

By decentralizing the process of decision making, you mitigate the risks, since you are allowing entities to tackle a problem in parallel, and if one of them turns out to be wrong, the fallout is localized. Whereas in a centralized system, one bad decision can spell trouble for everyone.

View on HN

> It's not like every central institution can marshal expertise on every local issue.

True, but maybe we disagree on how easy it is for every locality to acquire expertise. Plenty of local governments in the United States are already stretched thin fulfilling the simpler functions of policing, schooling, and basic infrastructure. Adding a staff epidemiologist seems very hard for such places. The same proviso applies to many poorer states and nations. That is why I think it makes most sense to direct effort toward having a competent central authority.

> By decentralizing the process of decision making, you mitigate the risks, since you are allowing entities to tackle a problem in parallel, and if one of them turns out to be wrong, the fallout is localized. Whereas in a centralized system, one bad decision can spell trouble for everyone.

On the flip side, decentralizing makes coordination harder. In coarse terms, there are game-theoretic considerations in an epidemic. For example, if localities are purely self-interested, an unaffected region may attempt to quarantine and keep all of its medical infrastructure ready for itself rather than lend it to nearby affected regions. In such cases a coordinating mechanism can be extremely helpful.

I'm not arguing that one kind of organization is always better or worse. To me it just seems complicated.

> On the flip side, decentralizing makes coordination harder.

Perhaps technology could help in this regard, at least to some degree. Bitcoin has proven coordination in a low trust environment is possible. Imagine if we could make coordination easier in other industries too.

> I'm not arguing that one kind of organization is always better or worse. To me it just seems complicated.

I agree, it is indeed complicated. But we need to start seriously considering concepts such as scale, systems thinking with its centralization/decentralization paradigms, and incomplete knowledge/risks if we want to design better institutions. Then we can thinker with turning the knobs one way or the other.

> For example, if localities are purely self-interested, an unaffected region may attempt to quarantine and keep all of its medical infrastructure ready for itself rather than lend it to nearby affected regions. In such cases a coordinating mechanism can be extremely helpful.

Reality is already ahead of you. [1] Switzerland is currently having trouble receiving masks and other medical supplies that were commissioned out of state:

Being landlocked, goods bound for Switzerland necessarily have to pass through a different country first, in this case Germany. Germany has a decree by which Corona-related medical supplies must not leave the country - and right now, this even applies to goods destined to Switzerland that are just arriving at a german port.

A fully "decentralised" world full of self-interested microstates would see a lot more of this stuff.

[1] https://www.thelocal.de/20200312/mask-hysteria

Reply on news.ycombinator.com